battery
AGENTS.md0 starsChanged 2 months agoMITView on GitHub
North star: solve real user pain (session amnesia, context rot, tool switch, machine migration) — not feature sprawl. Do not stack overlapping skills in one turn. Do not skip ADRs/roadmap for greenfield features. No feature merges without a scorecard row in docs/metrics/IMPROVEMENT_SCORECARD.md: baseline metric, post-change metric, target, pass/fail. Run uv run battery eval --real and/or --stress when retrieval is touched. Installed for this repo: product-manager-toolkit, roadmap-communicator, architecture-decision-records, hybrid-search-implementation, context-engineering, spec-driven-development, brainstorming, domain-modeling, implement, codebase-design, documentation-and-adrs, and others. Global Matt skills…
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 33,315 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository14 other files it gives its agents.
- architecture-decision-recordsSkill
- brainstormingSkill
- codebase-designSkill
- context-engineeringSkill
- documentation-and-adrsSkill
- domain-modelingSkill
- hybrid-search-implementationSkill
- implementSkill
- mermaid-expertSkill
- product-manager-toolkitSkill
- product-strategistSkill
- roadmap-communicatorSkill
- spec-driven-developmentSkill
- vector-index-tuningSkill
The file
We split this AGENTS.md into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 49
Team: commit BATTERY.md to git
≈ “Common” is an estimate: at least 1 in 5 of the 42,283 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Battery — Agent Working Agreement
North star: solve real user pain (session amnesia, context rot, tool switch, machine migration) — not feature sprawl.
Phase discipline (one phase → one skill)
| Phase | Read first | Skill |
| Problem / priority | docs/roadmap/PRODUCT_ROADMAP.md | .agents/skills/product-manager-toolkit |
| Shape / explore | docs/adr/, docs/design/ | .agents/skills/brainstorming |
| Spec before code | New ADR draft | .agents/skills/spec-driven-development |
| Domain / terms | ADR + this file | .agents/skills/domain-modeling |
| Module design | Similar modules in src/battery/ | .agents/skills/codebase-design |
| Implement | Approved ADR + ticket | .agents/skills/implement |
| Debug / regressions | Failing test or eval | ~/.agents/skills/diagnose (global) |
| Review | Before merge | ~/.agents/skills/code-review (global) |
| Context setup | Session start | .agents/skills/context-engineering |
Do not stack overlapping skills in one turn. Do not skip ADRs/roadmap for greenfield features.
Scratchpad (thinking out loud)
- Active hypotheses:
docs/scratchpad/ACTIVE.md - Promote stable decisions → ADR (
docs/adr/) - Promote shipped work → roadmap (
docs/roadmap/PRODUCT_ROADMAP.md)
Canonical docs (always ground here)
docs/adr/README.md— architecture decisionsdocs/roadmap/PRODUCT_ROADMAP.md— prioritiesdocs/metrics/IMPROVEMENT_SCORECARD.md— baseline → delta → pass/faildocs/design/battery-master-spec.md— system specdocs/design/cross-verification-audit.md— assumption ledger
Ship gate
No feature merges without a scorecard row in docs/metrics/IMPROVEMENT_SCORECARD.md: baseline metric, post-change metric, target, pass/fail. Run uv run battery eval --real and/or --stress when retrieval is touched.
Memory model (ADR-0007)
| Tier | Categories | Inject by default? |
| Stable semantic | rule, decision, preference, general | Yes |
| Event episodic | episodic | No — handoff/search |
| Raw chat (future) | chunked episodic | Opt-in only |
Portability
- Team: commit
BATTERY.mdto git - Individual:
battery profile export/battery profile import - Models: ONNX weights not in bundle (~90MB re-download)
Project skills (.agents/skills/)
Installed for this repo: product-manager-toolkit, roadmap-communicator, architecture-decision-records, hybrid-search-implementation, context-engineering, spec-driven-development, brainstorming, domain-modeling, implement, codebase-design, documentation-and-adrs, and others.
Global Matt skills remain in ~/.agents/skills/ per dev-phase-skills cursor rule.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

